Professional Documents
Culture Documents
research-article2014
MAC0010.1177/0020294014534205Tech Talk: (2) Process Control BasicsTech Talk: (2) Process Control Basics
Contributed Paper
Tech Talk
Measurement and Control
2014, Vol. 47(5) 146–152
© The Institute of Measurement
and Control 2014
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0020294014534205
An Introduction
In conducting over many years the C&I engineers need a good working both the fundamentals of measurement
assessment of Control & Instrumentation knowledge of many subjects, including and control and its practical application.
(C&I) engineers applying for professional the wide range of field instruments and To this end, experienced engineers and
registration, it is apparent that significant control systems available in a fast manufacturing companies are cordially
gaps often existed in their fundamental changing marketplace, materials invited to share their knowledge for the
technical and application knowledge. selection, piping mechanics, fluid physics, greater good by submitting for
Notwithstanding the normal variations industry-specific measurement, publication, further papers that will be of
in the ability of individuals, the academic protection and control techniques/best interest to practising C&I design
route chosen and the subsequent practice, functional safety, hazardous engineers and technicians. Publication
training offered by employers will play a area work, control systems architecture/ guidelines are available on request from
major part in the development of a well- application, data communications, publications@instmc.org.uk
rounded engineer. installation design, engineering standards/ This first paper in the series addresses
Professionals following the craft codes of practice, inspection and testing, the basics of measurement. The
apprentice/Higher National Certificate contractual awareness and much more. co-authors are David W Otterson CEng,
(HNC) route through technician training and Large companies with accredited FInstMC, and John E Edwards CEng,
who are able to obtain employment in an graduate training programmes will usually FIChemE, who has published several
engineering design capacity will usually make use of both in-house mentoring papers and books on process simulation,
possess good practical C&I application and external training, whereas smaller instrumentation and control.
skills. In contrast, their university graduate organisations are often much more reliant
counterparts, although better equipped on the haphazard self-learning process. David W Otterson
mathematically, may struggle in their Tech Talk is a series of papers CEng FInstMC
formative years with the broad spectrum of designed as a ‘pull out’ reference library. Institute of Measurement and Control
practical application knowledge required. The aim is to disseminate knowledge of Billingham, UK
I. Introduction time and capacity. Dead time is the to be applied. Flow and liquid pressure
interval, after the application of an action, loops exhibit low capacity and no dead
Stable control is essential for the safe and during which no response is observed time, giving fast and noisy characteristics
efficient operation of industrial processes. and capacity is a location where mass or requiring a low gain (wide proportional
Understanding the basics of process energy can be stored. These dynamic band) response with reset (integral
control is essential for the instrument elements determine the way control action) to maintain a desired value.
engineer or technician to effectively loops behave in terms of speed of Gas pressure loops usually exhibit high
design, maintain and tune instrumented response and stability. capacity and no dead time, with slow
control loops with confidence. Each process parameter exhibits characteristics, requiring a high gain
All processes contain, in various unique properties that determine the (narrow proportional band) response to
proportions, the dynamic elements dead ease of control and the control strategy maintain a desired value. Liquid-level loops
exhibit high or low capacity and variable II. Control Loop Presentation Symbols are defined using international
dead time, giving fast or slow, noisy standards according to Instrument
(boiling) and lag characteristics requiring A loop schematic or process control Society of America (ISA) or British
medium gain and in certain applications diagram is a single line diagram used Standards (BS):
(boiler level control) reset response to to identify the basic specifications
maintain a desired value. Temperature and structure of an instrument control BS 1646-3:1984 Symbolic
loops can exhibit high or low capacity and loop and is developed from the representation for process
dead time, giving fast or slow and lag Process & Instrumentation Diagrams measurement control functions and
characteristics requiring variable gain, (P&ID). A completed loop schematic instrumentation.
reset and anticipatory (derivative) response diagram should provide all the details
to maintain a desired value. necessary to enable the loop American National Standards Institute
The process plant design and the interconnection drawings to be (ANSI)/ISA-5.1-2009 Instrumentation
control philosophy should be considered prepared correctly. Symbols for Identification
throughout a project design process. For The main features include unique
example, if a processing unit takes a long identification and specification of all Each element within the loop is
time to achieve stable and optimal components in the loop, the location identified by a symbol containing the
operating conditions, the feeds should of the components, signal unique identification (Tag Number). The
be held constant by the use of surge interconnection type, power and air symbol used defines the function,
tanks in which the level is allowed to float supply type and source. Cable, wiring accessibility and the location as shown in
between acceptable limits. and termination details are excluded. the Table 1.
Discrete instruments
FIC FI FX FT
101 101 101 101
Computer function
Programmable logic
Typical tag numbers are developed as The instrument loop number should components with respect to each other.
follows: convey as much information as possible, The line weight and format represents
such as plant area designation and different signal types.
FIC 101 Instrument identification or assigned blocks of numbers for specific
Tag Number; functions. When scheduled, fields are
III. Control Loops
F 101 Loop identification; function provided to the right of the symbol to
descriptor (F-Flow, T-Temperature, describe the component function, device An open loop (Figure 1) involves an
P-Pressure, L-Level) manufacturer/type specification number operator manually setting a signal to
and calibrated range. Signal levels and manipulate a control device, such as a
101 Loop number; power or pneumatic supplies are also control valve or a variable speed drive to
FIC Functional identifier (I-Indication, shown. On the P&ID, the component manipulate a process variable. The result of
C-Control, T-Transmitter, symbols are interconnected to show the this action may be monitored by a separate
X-Miscellaneous); relationship and functionality of the measurement and indication loop.
Integral action time Ti, min(s), longest in the controller driving to its maximum or unanticipated load changes. Td usually
time (or off); minimum output: set in the range Ti/6 < Td < Ti/2 is
considered reasonable.
Derivative action time Td, min(s),
shortest time (or off). Use a low Ti in the range 0.05–2 min
on fast and noisy processes: flow and D. Integral action time saturation
A wide proportional band should be liquid pressure. (wind up)
used on noisy signal control loops, for When a sustained error is maintained
Use a high Ti in the range 2–120 min
example, liquid flow, liquid pressure and between the set point and the measured
on slow processes: temperature,
level; note the controller gain is 100/P. variable on a controller with Ti, the output
vapour pressure and composition
Integral action times (Ti) are used to will drive off scale (valve full open or
control.
eliminate error; short times can make the closed). This condition is typical of
loop less stable. Derivative is never used heating and cooling a batch reactor to a
It is used on critical level control
on noisy signals as Td operates on rate of desired temperature and composition
applications such as boiler drum level
change of error de/dt, where rapid endpoint control. Controllers
and distillation column reboiler level
changes in error will result in large rapid incorporating Proportional and Derivative
control, where settings are application
changes in controller output leading to modes only do not experience this
specific. It is not used on liquid level
control valve slamming and damage. problem. Modern controller designs
control where a specific control point is
not required, thus allowing for steady employ nonlinear techniques to prevent
A. Proportional band (P) flow conditions, for example, surge tanks the integrator output becoming too high.
The proportional band setting between process units. The use of Ti is
determines the magnitude of the often unnecessary on gas pressure E. Controller tuning methods
controller output change as a result of control.
There are many practical and theoretical
the error between the process
approaches to controller tuning. Modern
measurement and the controller set C. Derivative action time (Td)
controllers are provided with auto-tune
point. There is no error at one specific
Derivative Action Time changes the facilities which should be applied with
process load condition:
controller output at a rate proportional to caution on new systems. A long
the rate of change of the error de/dt and established procedure due to Ziegler and
Use a narrow P in the range 1%–50%
provides an anticipatory characteristic. Nichols known as the Ultimate Sensitivity
on slow processes: Gas pressure,
The controller action can be likened to Tuning Method will produce conservative
temperature and vapour pressure
that of an experienced operator adjusting parameters allowing for further final fine
control
a heating valve to achieve a set value as tuning.
fast as possible. Not normally used on Set the control modes to their
noisy processes such as Flow, Liquid minimum effective values: that is,
B. Integral action time (Ti)
Pressure and Liquid Level. maximum P, maximum Ti and minimum
Integral Action Time changes the The fast rates of measurement change Td (or off). Reduce P while making small
controller output signal at a rate can lead to control valve slamming and set point changes in both directions until
proportional to the magnitude of the error damage. Td is typically used on the measured variable begins to oscillate
until the error is eliminated. It should Temperature, Vapour Pressure and at constant amplitude, providing a trend
never be set so fast that the resulting Composition and should be set reading of the proportional band Pu and
load change exceeds the process conservatively (short Td) to protect the oscillation period Tu. Initial controller
response characteristics which will result against control instability resulting from settings are determined from Table 2.
P only 2 Pu
Figure 3.
This procedure can be used as a tuning parameter settings (Figure 3(b)) F. Interpretation of Figure 3 trends
‘starting point’ for tuning control systems when controlling level in a surge tank (note that derivative mode is not
on batch or continuous processes, Tu with a controller set point of 2 m. The used due to the noisy signal)
being established at a stable condition. initial inlet flow is set at 4000 kg/h with a
The following plot (Figure 3(a)) shows the step change to 16,000 kg/h after 15 min P only set at 25%. Initial error is 0.1 m.
process simulation to study the effect of and then back to 4000 kg/h after 50 min. There is a specific flow condition that will
Funding
This research received no specific grant from
any funding agency in the public, commercial
or not-for-profit sectors.